JP2816691B2 - Optical switch assembly - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates generally to optical switches, and more particularly to optical switch assemblies for exchanging optical signals in free space.

2. Description of the Related Art A 2 × 2 directional coupler is used to switch an optical signal in an optical waveguide. A problem of such a waveguide directional coupler is a power loss when the waveguide is directionally coupled at input and output. As a result, the size of the switch matrix including these directional couplers is restricted. One such methodical coupler is a lithium niobium crossbar coupler, which has a long physical structure. As a result, a large switch matrix containing this crossbar coupler requires physical bending, which increases power loss.

Optical crossbar switches are also known that switch an optical signal in free space from an arbitrary position in one column to a selected position in a single row without using a waveguide. A two-dimensional array of selectors / decoders is typically provided between the two lenses to implement this crossbar optical switch. However, this optical switch switches the optical signal in one dimension. That is, it is limited to switching from a single column to a single row and vice versa.

A single stage of these prior art optical crossbar switches can implement time division optical switching such as reordering serial bits received on an optical fiber. Using one stage of such an optical switch, it is also possible to realize space division switching in which a serial bit stream is switched from one input fiber to a selected output fiber. Furthermore, these optical switches can be interconnected in several stages to perform space-time exchange. However, the problem with these optical crossbar switches is that they cannot be combined to perform a combined exchange of space and time division in a single stage.

SUMMARY OF THE INVENTION In the illustrated embodiment of the present invention, an optical information signal is output from a selected row of an input array by a free space optical switch assembly having a two-dimensional input array, a distributor, and a two-dimensional output array. The disadvantages associated with the above-described problem are resolved by switching to the selected row. The switch also includes an input system for converting a serial bit sequence of the temporally separated optical information signals into spatially separated parallel bit information stored in the input array. As a result, a significant technological advance can be made in that a single switching stage can perform both spatial and temporal switching of optical signals.
Furthermore, the optical switch assembly can advantageously exchange two-dimensional optical signals, such as rows of an array of light sources. The distributor of the optical switch of the present invention spatially distributes the electric field pattern. As a result, for example, an optical signal from any of a plurality of input positions is converted into an output electric field pattern, and all of the plurality of output positions are output. In the light of
The magnitude of the output electric field pattern can be converted to a Fourier transform of the input electric field pattern. Distributor
For example, the optical signals of the individual columns from the selected row of the array of input locations are distributed centered about their optical axis and the center of the corresponding columns of the output array, and each position of each corresponding output column is It can be illuminated. The same result is obtained regardless of the vertical position of the optical signal in the input train.
Similar results are obtained for each optical signal emitted from the other column of the row of the selected input array.

Also included in the optical switch assembly are encoders provided at each input location for selectively emitting an optical input signal from the selected input location. The encoder controls the order in which selected optical signals pass through the distributor. In the illustrated embodiment, where the input locations are arranged in rows and columns, the encoder selectively emits one row of the spatially separated optical signal at a time to provide light to the distributor.

An input system of the switch, for example, an array having a row of optical input shift registers that converts a temporally separated serial bit sequence of optical information signals from a plurality of input optical fibers into a spatially separated information bit pattern; An input storage array having a row for storing a spatially separated information bit pattern from each shift register row is included. For example, serial, temporally separated information from an optical fiber may be simultaneously applied to one end of a group of rows of shift registers. Only one shift register in each time slot or group of time periods receives a temporally separated information signal so that the spatially separated information can be shifted over the entire length of the row. That is, when M words arrive at the transmitter from the serial data stream, each word is shifted into a separate one of the M shift registers. During each successive time slot, another individual shift register row is energized to receive and shift the information for that time slot. After information has been provided to all shift register rows in the group, the spatially separated information bits of all rows are transferred as a whole in parallel and provided to the input storage array.

In addition to transforming temporally separated information into spatially separated information, another advantage of this input system is that the input storage array storing the information can be, for example, one frame of spatially separated information. To the encoder array,
While giving it selectively to distributors,
The shift register array can be provided with a continuous time frame of information temporally separated from the optical input fiber.

An important consequence of this spatio-temporal transformation is that only shift register arrays may be able to operate on the data rate of the temporally separated information. The input storage array, decoder and distributor can operate at lower operating speeds than temporal information.

In order to improve the signal-to-noise ratio of the optical signal distributed to the output locations, the decoder arrangement periodically positions the optical devices, for example, by arranging the input arrangement in rows and columns, and From the selected device or row of devices. The signal-to-noise ratio of the distributed optical output signals of the columns of the output array reduces the distance between the encoder devices in the columns of the input array, but also by controlling the shape of the emission area of the encoder devices. Significant improvement over switches. This is a notable advantage of the optical switch of the present invention, for example, since the contrast ratio of a bistable encoder device can be as low as 1.5 to 2.0. Even if there is no infinite contrast ratio between the bistable conducting state and the non-conducting state of the optical element, the noise signal of the light of each element in the input string is spatially arranged in the output array first with the selected optical information. Be distributed. The encoder advantageously controls the contribution of each optical noise signal to the input electric field pattern such that its Fourier transform is reduced to a predetermined output electric field pattern distributed over the entire height of each output array column. . It is known that the intensity distribution of the input spot to the distributor is defined by the size and shape of the modulator in the encoder, and the effect of the Fourier transform performed by the distributor is also very well known, so the output The intensity pattern generated by the distributor in the device array can be easily predetermined. If the contribution of each input noise signal is essentially uniform, the signal-to-noise ratio at a predetermined location in the output sequence is greatly improved, and the decoder elements may be advantageously located there. .

To further improve the signal-to-noise ratio at predetermined locations in the output array, the optical switch assembly further has a Gaussian distribution for the encoder array, for the entire height of the array of encoder arrays, and Means are provided for providing an optical power signal having a predetermined magnitude such that it has a uniform distribution over the entire length of the row. The resulting Fourier transform of the optical noise signal and the optical information signal further improves the signal-to-noise ratio at selected locations along the height of the output train. More specifically, the de-energized spots of the input device array may produce some undesired output power, and because the de-energized spots are on regulated, regularly spaced, these de-energized spots. The power from the energizing spot will be directed to discrete locations in the output device array by the shadow of the Fourier transform function. In particular, the Fourier transform of any periodic function is itself a periodic function. The undesirable power from the activation spot represents noise in the distributor system, and thus has the advantage that most of the noise power is directed discretely in the output device array. If the output detector is located at a distance from the discrete location to which the noise power is directed, the receiver can operate at very low noise levels. This is why "the signal-to-noise ratio at selected locations along the height of the output train shows a further improvement".

A decoder array having a plurality of light storage elements provided at the output of the switch to complete the switching function selectively stores the spatially distributed optical signals passing through the distributor. This causes the optical signal from the input array location to be switched to any one or more of the output array locations. In the switch shown, the decoder storage element also has the advantage of taking advantage of the improved signal-to-noise ratio associated with the Fourier transform of the periodically stored light source.
Columns and rows having the same number are arranged and arranged in the same manner as the encoder array. The information stored in one or more of the rows of the one or more selected output arrays is then provided to an optical output system for further space-time conversion and distribution to a plurality of output filters. By arranging the decoder storage elements in rows and columns, the broadcast function can be advantageously implemented by having one or more rows of the array correspond to individual communication channels.

In the illustrated optical switch assembly, the distributor has an optical axis about which axis the optical signal propagates from any of the input locations. As an example, this includes a cylindrical convergent lens that produces a Fourier transform at the focal plane of the lens.

In order to illuminate all of the output locations, with adjacent input and output arrays, the distributor also includes magnification means such as, for example, a cylindrical dispersion lens for magnifying the vertically passing optical signal.

The distributor of the optical switch further includes imaging means, such as a cylindrical converging lens for limiting the dispersion of the optical signal in the horizontal direction. This prevents interference between adjacent columns and reduces optical power loss.

Another advantage of the optical switch of the present invention is that it converts the spatially separated information into a serial bit stream of temporally separated optical information and sends it to optical information for delivery to a plurality of output fibers. To store the light output system. As with the optical input system, the output system shifts the information stored in the rows of the output memory array by storing the selected row of information received from the decoder array in rows and the information stored in the rows of the output memory array. And a plurality of output shift register rows for sending to the output optical fiber as a serial bit string of the separated optical information signals.

BRIEF DESCRIPTION OF THE DRAWINGS The present invention will be more fully understood from the following detailed description when read in conjunction with the following drawings.

DETAILED DESCRIPTION FIG. 1 shows an optical switch 100 for performing space-division exchange in a single-stage switch for performing exchange in the optical domain without using a waveguide in free space. Basically, the switch consists of an input system 101, an encoder array 102, a distributor 103, a decoder array 104 and an output system.
105, which is under the control of the switch control circuit 110. The input system 101 includes an input shift register 131 having a row of shift registers and an array of storage elements 132 facing the row of shift registers and having a corresponding position relative to the row of shift registers. The input system spatially separates the temporally separated information represented by a pair of serial bit streams of complementary optical information signals received from each pair of input optical fibers 150 (1) -150 (N). And store the spatially separated information in a row of the shift register during a given time period (temporally separated information refers to a time division serial data stream. Since the input to the system is assumed to be a collection of optical fibers, two fibers are required for each of the serial data streams, each pair of fibers carrying a single time-division multiplexed differential data stream and a logical For "1", fiber # 1 has optical power and fiber # 2 has no optical power, and for logical "0", fiber # 1 has no optical power. # 2 has optical power. Thus, each fiber carries a binary phase correction signal to its associated fiber.) This time period is called a time frame containing a plurality of time slots. The information or data in each time slot typically includes a sample of data from the communication channel. Each bit pair of the complementary optical information signal represents a single bit of binary information. Information is entered into the shift register such that each time slot of data is stored in a single shift register row. At the end of each time frame, all spatially separated information in the row of the shift register is transferred and stored in the corresponding row of the input storage array to select the encoder during the next time frame. During this operation, the information of another time frame is given to the shift register unit.

Under the control of the switch control circuit 110, the encoder array 102 selects an optical signal representing the information stored in the input storage array row by row, and only one row of the spatially separated optical information signals at a time. Propagate through the distributor. Distributor 103 spatially distributes each pair of optical signals so that the optical signals are emitted from selected encoder rows and run through the height of each corresponding column of decoder array 104. As a result, each row of the decoder array has access to the information stored in any row of the input storage array selected by the encoder array. The decoder arrangement has the same column numbers as the input encoder arrangement, both of which are in the switch control circuit 110.
Under the control of. The decoder arrangement selects one or more of the rows and transfers the information to an output system for storage there. If the decoder array has at least the same number of rows as the encoder array, the switch operates as a completely non-blocking switch.

Like the input system 101, the output system 105 includes an output shift register unit 129 with an array of storage elements 133 and a row of shift registers. During a given time frame, the information in each row of the input storage array selected by the encoder array is spatially switched to the row of the output storage array selected by the decoder array. This allows information in any row of the input storage array to be selectively transferred to one or more of the rows of the output storage array. At the end of the time frame, all the spatially separated information in the output storage array is transferred to an output shift register unit where it is stored and converted into time separated information.
The information in the spatially separated format is again represented as a pair of complementary optical information signals, and a plurality of optical output fiber pairs.
160 (1) to 160 (N). As a result, the optical switch 100 switches information from any input fiber pair to any output fiber pair and further converts serial information, such as the data slot of the input fiber pair, to any other time slot of the time frame. To perform a time division exchange by switching to In addition, the optical switch simultaneously time-divisions and simultaneously switches information of any time slot of a given time frame from one input fiber pair to the same time slot of any one or more output fiber pairs. Perform a space division exchange.

FIG. 2 shows a detailed view of the input system 101. The input system is an optical input interface device 106, an optical input shift register device 13 having shift register columns 131R1-131RP.
1.Intermediate interface device 135 and input storage element array
Contains 132. The input shift register device 131 is connected to the input optical fiber 150 via the optical input interface device 106.
(1) Receive temporally separated information represented by serial bit pairs of complementary optical information signals received from each pair of -150 (N).

As shown in FIG. 2 and shown in more detail in FIG. 12, input shift register 131 has rows 144R1-144RP and columns 1
Optical storage elements 144 assigned as 44C1-144C8, and corresponding arrays of holograms 139 and input shift registers 131
Includes an imaging system 138 located therebetween forming R1-131RP. Similar to an electrical shift register having master and slave storage at each bit position, rows 131R1-131RP of each shift register have master and slave optical storage elements at each bit position. Assuming that each shift register row stores and shifts eight bits of information, each shift register row includes 16 optical storage elements, which are provided in 16 correspondingly arranged hologram arrays 139. Optically interconnected by the provided hologram.

As shown in FIG. 12, the shift register 134R1 comprises a master and slave optical storage element pair 144 (R1, C1) M, S-139.
(R1, C8) Contains M and S. Row 139R1 of hologram array
Consists of similar reflection-type hologram pairs 139 (R1, C1) M, S-139 (R1, C8) M, S having a diffraction efficiency of 100% or less, and the optical storage elements 144 (R1, C1) MS-144, respectively. (R1, C8) M and S are placed relative to each other. Row 132R1 of the input storage array is located opposite optical storage elements 132 (R1, C1) -132 (R1, C8) M.

As shown in FIG. 2, each input fiber pair is uniquely associated with a plurality of input shift register rows.
For example, input fiber pair 150 (1) is associated with input shift register rows 131R1-131R3. With this arrangement, the time slots of the data, represented by the time slots of the complementary optical information bit signal pairs received serially and received on a given input fiber pair, are shifted by individual shifts associated with the input fiber pairs. Will be stored in the register row. The number of rows associated with each input fiber pair depends on the data rate of the received optical signal and the input storage array.
It depends on how often each of the 132 rows must be distributed to the rest of the switch.

The optical input interface device 106 includes an optically transparent spacer material 107 and a plurality of transmission holograms 10 for providing a pair of optical information signals received from an input fiber pair to a storage element.
8 and a beam splitter 109 for providing an optical power signal, such as a clock synchronized timing control signal, to each of the storage elements of the input shift register device 131. Spatial light modulator such as the well-known Semetex Cytomod modulator
140 is a light source (not shown) under the control of the switch control circuit 110
Emits clocked optical timing signals 163-165 in response to the optical bias beam 151 from Upon receiving a control signal from the control circuit 110 switched via the bus 170, the modulator controls the timing signal as shown in FIG.
Emit 163-165 to store, shift, and transfer information to a selected input shift register row from a given optical fiber.

The hologram fringe pattern pairs 108 (1) -108 (N) of the transmission hologram 108 are the respective input fiber pairs 15
The serial optical information signal pair from each of the 0 (1) -150 (N) is split and redirected to a row uniquely associated with the input fiber. Because each of the rows associated with a particular input fiber receives an optical information signal pair simultaneously, during a given time slot only one of the rows of the input shift register associated with that fiber will be active Receive and store information represented by the information signal pair. The hologram pairs are formed in a known manner and split the optical signal from a given input fiber pair and change direction so that at the end of each storage element row, it goes to the optical storage element in column 144C1. I do. Each of the input fibers is fixed to the spacer 107 in a known manner. The thickness of the spacer material 107 is selected to be long enough that the optical information signal from the input fiber pair can be dispersed enough to illuminate the entire fringe pattern pair of the hologram associated with the input fiber pair. I have. As shown in FIG.
The optical input carrier facility 180 containing (1) -150 (N) provides a known timing and synchronization signal to the switch control circuit 110 via the bus 166.

With the well-known polarization type beam splitter 109, an optical information signal can be provided from the input fiber pair 150 (1) -150 (N) to the shift register storage elements in column 144C1 with little loss. The beam splitter can also direct optical signals, such as optical timing signals 163-165 for the clock, to each of the storage elements of the input register device 131. The optical timing control signals 163-165 are coherent or incoherent light having a uniform electric field distribution, depending on the type of optical element used at 144 as a storage element, which is provided to each of the elements of the array 131.

Each of the optical storage elements of the shift register device 131 stores a bit of binary information represented by a complementary optical information signal received on one of the input optical fiber pairs 150 (1) -150 (N). An optical device suitable for use as a single-input, single-output optical storage device is the self-electro-optic device described in U.S. Pat. No. 4,546,244. These elements are interconnected by light to form a shift register and store bits of serially received optical information as rows of a storage array. Self-Electro-Optical Device (SEED) is available from HS Hinton US Pat. No. 4,764,890 and HS Hinton and DA
Interconnected and controlled by light as described in B. Miller U.S. Pat. No. 4,764,889. These patents describe a method of interconnecting optical shift registers by light to form an optical shift register for use as input shift register device 131.

Suitable devices for use as input shift register device 131 are HS Hinton, AL Lentin and DA
B. Miller is described in U.S. Patent Nos. 4,715,378 and 4,754,132. These documents describe a self-electro-optical effect device in which two complementary, symmetric optical signals are stored in these devices when the ratio of the two optical signals incident thereon exceeds a predetermined threshold. Says. The information stored therein is read by applying optical signal power to the two quantum wells of the device simultaneously. At this time, two complementary and symmetric optical output signals are emitted from the device representing the information stored therein. Other known optical devices may be used to store the information represented by the binary optical signal.

During a given time slot, the input hologram pair
Each of 108 (1) -108 (N) is a row of shift register devices.
Providing a complementary pair of serial optical information signals from each pair of input fibers to the 131C1, which then shift one bit at a time and select the shift when the binary for that time slot is received. Shifted to register row. Thereafter, the data of the next time slot is stored in the row of another shift register associated with the fiber pair. The data of one time slot is stored in the row associated with each fiber pair until all rows have been stored, and the time slot spatial information of the information associated with a given time frame will be separated. The number of rows associated with each input fiber pair is determined by the number of information time slots per time frame.

FIG. 13 illustrates a timing diagram for the time t of the idealized optical information signal 1301 on the input fiber pair 150 (N) and the idealized optical timing control signal 162-165 provided to the optical system 101. . Time frame F data 13
Only the last time slot of 02 is referred to as data or information, and shows only a portion of the first data word of time frame (F + 1). Each word of the data of the optical information signal pair 1301 includes 8-bit information B1-B8.

In FIG. 13, the input timing control signals 162-165
Is referenced and synchronized with an input reference clock signal on timing bus 166, which provides information about the boundaries of incoming information bits. In response to the reference clock signal, the switch control circuit 110 controls the change of the optical timing control signal in a known manner. For purposes of explanation, assume that each time frame has three time slots or words. The 8-bit B1-B8 data word arriving sequentially on the input fiber 150 (N) is shifted and stored in the last row 131PR of the input shift register device 131.

In order to shift the last word 1302 of time frame F to the last input shift register row 131RP, the electrically controlled spatial light modulator 140 uses optical timing control with respect to time t as shown in FIG. Emit signals 163-165. Spatial modulator 140 alternately applies optical timing control signals 163 and 164 to the third of each plurality of rows associated with the input fiber pair.
Are alternately applied to the rows. Optical timing signal 165 remains high for all rows of input shift register device 131 that were not selected to provide a signal during this time slot. Optical timing signals 163 and 164 pulse at alternating high (1) and low (0) power levels for the last shift register row 131RP. Optical timing signal 163 is directed to the master flip-flop storage element at row 131PR, while optical timing signal 164 is directed to the slave flip-flop storage element at row 131PR.

At time to, the leading edge of the optical signal 1301 representing the first bit B1 is stabilized by the input fiber 150 (N), and each of the master flip-flop storage elements in the last three rows of column 144C1 of the input shift register storage array 144 Hologram 108
When distributed by (N), the timing control signal 163 incident on all master flip-flop storage elements in row 144RP changes from a high power level (1) to a low power level (0), thereby causing data to be lost. The data is stored in the first master flip-flop storage element 144 (PR, C1) M of the last shift register row 131RP. Similarly,
Spatial light modulator 140 provides timing control signals 163 and 164 to selected input shift register rows associated with each of the other input fiber pairs.

First master flip-flop storage element 144 (PR, C1)
After the data B1 is stored in M, the timing control signal 163 that enters the master flip-flop storage element at time t1
Changes to a high power level (1) and is applied to the slave flip-flop storage elements in the same row.
4 changes from a high power level (1) to a low level. First
When the optical signal of the optical power level pair is incident on the master flip-flop including the storage element in the flip-flop of
The data stored in each of them propagates out of the hologram row 139RP as a pair of complementary optical information signals, and the hologram row changes the direction of this signal and supplies it to the adjacent slave flip-flop element in row 144RP. . Last row 144R
Since the timing control signal 164 is at a low power level in the P slave flip-flop element, each piece of information represented by the complementary information signal pair from the master flip-flop element is stored in an adjacent slave flip-flop element. The portion of each optical information signal pair transmitted by the hologram in row 139RP has no effect on the input storage array 132. This is because while shifting input data through the rows of the input shift register, the timing control signal incident on the entire input storage array is at a high level. As a result, the data bit B1 is shifted from the first master flip-flop storage element 144 (RP, C1) M to the first slave flip-flop storage element 144 (PR, C1) S.

At time t2, the timing control signal 164 incident on the slave element changes to a high level, and the complementary information signal pair 130
1 takes a complementary signal level representing the data bit B2 between times t2 and t3.

At time t3, when the incoming bit B2, which appears on the input fiber pair 150 (N), becomes stable, the control signal 163-165 changes as described above, resulting in the data bit
B2 and B1 are stored in the continuous slave flip-flop elements 144 (RP, C1) and 144 (RP, C2), respectively. This sequence of light control signals 163 and 164
This is repeated until B1-B7 has been stored in successive slave flip-flop elements in row 144RP.

At time t5, the information signal pair 1301 representing the bit B8 on the input fiber pair 150 (N) becomes stable, and the hologram 108
When directed to storage element 144 (PR, C1) M by (N), timing control signal 163 again changes low at the master flip-flop element on row 144RP, thereby causing bit B8 to change to master flip-flop storage element. 144 (RP, C
1) Energize to memorize in M. At the same time, the information bits B1-B7 propagate from the slave flip-flop element to the next succeeding master flip-flop element via the optical information signal pair. At time t6, the timing control signal 163
Returns to the high power level, which causes bits B1-B8 to be stored in the master flip-flop element of storage row 144RP. Since this is the last word of frame F, ie, the time slot, the clock light source 143 shown in FIGS. 1 and 2 assumes a low power level at time t6 in all the flip-flops of the array of input storage elements, On the other hand, the timing control signals 163-165 operate the timing control signal 162 so as to take a high power level. This causes all 8-bit data words to be input from the storage elements of the master flip-flop in array 144 through a hologram array, each of which reflects a portion of the light and transmits the remaining portion of the optical information signal pair, to input storage array 132. Are transferred in parallel. The timing control signal 163 during this parallel data transfer
The reflected light does not affect the data stored in flip-flop storage element array 144 because -165 is at a high power level. This parallel transfer of data is completed at time t7, when signal 162 returns to the high power level. At time t8, time frame F ends and time frame (F + 1) starts. Prior to time t6, all P data words stored in the input storage element array during the previous time frame (F-1) are provided to the decoder array 104 via the encoder array 102 and the distributor 103. Will be.

The interface unit 135 is a known optical imaging system 136 and a beam splitter 137, which is a known polarized waveform. The imaging system 136 individually focuses the optical signal pairs emitted from each hologram of the hologram array 139 onto corresponding storage elements of the input storage array 132. Optical information signals are transmitted through the beam splitter 13 with minimal power loss.
Propagate through 7. Further, beam splitter 137 directs clocked optical timing control signal 162 from clock light source 143 to elements of input storage array 132. Switch control switch 110 controls clocked light source 143 through control bus 175.

As shown in FIG. 1, the optical information signal pair 192 (P, M)
Propagates from the input storage array 132 through the optical encoder interface device 111 to the encoder array 102. The encoder interface 111 includes an optical imaging system 112 and a beam splitter 113. Similar to imaging system 136, known imaging system 112 focuses light signal pairs from each storage element pair of input storage array 132 via beam splitter 113 on corresponding element pairs of encoder array 102. give. The polarized beam splitter 113 also provides a clocked optical bias beam 153 from a clocked coherent light source 114 to the encoder array in a known manner through a non-amorphous beamforming system 115. The switch control circuit 110 controls the clocked light source 114 through the control bus 168. A clocked coherent optical bias beam 152 is emitted from a known light source 114 and is an amorphous beamforming system 115.
Given to. Beam forming system is bias beam 15
An optical bias beam 153 having a uniform size of 2 and having a Gaussian distribution in one direction and a uniform distribution in a direction orthogonal thereto is formed. Beam splitter 113 changes the direction of the bias beam so that the Gaussian distribution is centered along the overall height of each row of encoder array 102. The uniform size of the bias beam 153 is distributed across the rows of the encoder array.

FIG. 3 illustrates the size of the optical bias beam 153 having a Gaussian distribution in the first direction and having a uniform size in a direction orthogonal to the first direction. Encoder array 1
By applying an optical bias beam having a Gaussian distribution to each column of 02, the signal-to-noise ratio (SNR) of the optical signal received by the decoder array 104 and the output system 105 is improved. To vary the signal-to-noise ratio of the light spatially distributed on the decoder array 104, the encoder array 102 may be provided with a bias beam having another distribution.

Encoder array 102 is a spatial light modulator that modulates information stored in input storage array 132 on a row-by-row basis. Switch control circuit 110 uses control bus 172 to scan a row of the encoder array in a known manner, energizing only one of the rows from the input storage array during any given row cycle period. Is stored. The rows of the encoder array are adapted to provide the optical information signals 192 from the input storage array 132 during each row cycle time.
(P, M), but only one row of the encoder stores information by switch control circuit 110 and then emits a coherent optical information signal pair to provide distributor system 103
Give to.

As shown partially in FIG. 1 and more particularly in FIG. 5, a preferred embodiment of the encoder arrangement 102 is
(C1) -102 (C8) and a plurality of three-state optical storage elements arranged in rows 102 (R1) -102 (RP). An optical device suitable for use as a three-state optical storage element in an encoder array is illustrated in FIG. 4, which is a modification of the electro-optical effect device described in U.S. Pat. No. 4,754,132. ing.

FIG. 4 shows the Intrinic (i)
A three-state symmetric self-electro-optical effect device 400 having photodetectors 401, 402 having regions of semiconductor quantum well regions 403 and 404, respectively. This device also has a photo detector 4
01 and 402 and a voltage source 405 include a conducting and non-conducting state switch 406 such as a transistor connected in series. When this is activated, the device will have complementary symmetrical and complementary high and low power levels in response to optical bias beams 153 simultaneously incident on the respective photodetectors 401 and 402 from terminals and Q. The optical information signal pair 193 (P, M) is emitted. Complementary optical information signal pairs 192 (P, P) from corresponding symmetric optical storage element pairs of input storage array 132
M), when the electronic switch 406 under the control of the switch control circuit 110 is energized and becomes conductive via the control bus 172, as described in the reference, switches the two states of the bias. Set and reset. When switch 406 is de-energized and non-conductive, each photodetector produces a low power level optical output signal pair when bias beam 153 is incident thereon. Device 400 operates as an optical SR latch, and the truth table for its logic states is as shown in Table A.

The truth table in Table A is the three-state SEED device in FIG.
It shows 400 physical operations. In particular, the device generates two light output signals, labeled Q and. The energizing input to the transistor is de-energized (En = "0")
When enabled, both outputs with Q are logic "0", indicating that very little optical power is being emitted from the two optical output ports of the device. However,
When the energizing input to the transistor is energized (En = "1"), the two light outputs (its spatial position depends on how the device was initialized by the inputs S and R)
A complementary output signal consisting of a light beam (logic "1") and a dark beam (logic "0") is generated.

In this way, the device acts like a clip-flop storage element whenever En "1" and
When it is "0", it acts like a tri-state (three-state) element (generating an unobservable output). This function is useful for implementing an encoder function in which only one row to the switch must be activated at any given time, and forcing all other rows to be deactivated. .
The row activated at the distributor at any given time is determined by a scheduling algorithm.

During each frame, the optical timing control signal 162 is at a high power level except for a short time before the end of the time frame as described above. While the signal 162 is at the high power level, the input storage array 132 provides the encoder array 102 with an optical output signal 192 (P,
M) is emitted from each storage element. As previously mentioned, switch control circuit 110 activates only selected rows of storage elements to store information from corresponding rows of storage elements of input storage array 132. Another requirement for storing only one row of information from the row storage array 132 of the encoder array 102 is that the coherent bias beam 153 is at a low power level and each energized encoder element is stored in the storage array 132. To respond to optical information signals from This results in an optical information signal pair representing the information stored in each bit of the energized encoder row.
193 (P, M) are emitted along with the low power level optical noise signal from each of the other elements in each column of the encoder array. The magnitude of each optical noise signal emitted from the train of encoders accumulates to approximate a Gaussian distribution. This is because the bias beam 153 has a Gaussian distribution. All optical noise and information signal pairs emitted from the encoder array are provided to distributor 103. During each time frame, switch control circuit 110 activates each decoder row once at a selected row cycle time to allow information signal pairs to pass through distributor 107.

Each of the optical information signals emitted from the storage elements, also referred to as pixels of the selected encoder row, is spatially distributed by the optical signal distributor 103 and mapped to a corresponding column of the decoder array 104. In particular, the distributor 103 spatially distributes the input electric field pattern of the optical signal from each element in the array of the encoder array, so that the magnitude of the output electric field pattern becomes a Fourier transform of the input electric field pattern, so that the individual An output electric field pattern large enough to illuminate all requests in the corresponding column. "Fourier Optics" by JW Goodman, published by McGraw-Hill
, The Fourier transform F of the complex function g of the two independent variables x and y is mathematically given by

Since the distributor 103 includes a cylindrical lens as described later, the necessary integration is performed only in the y direction. As a result, the Fourier transform results in one direction.

Here is g (y) = g (x , y), which is a pattern of electric field distribution was fixed to the value x ₁ of the given x.

Since the optical signals are spatially distributed over the entire height of the columns of the decoder array, the decoder array activates one or more selected rows during each row cycle time, similar to the encoder array. , Store the information represented by the spatially distributed optical signals. The decoder arrangement is then responsive to a clocked bias beam 154 to generate another signal pair 194 from each element representing the information stored therein.
(P, M) and output storage array 133 of output system 105
For the corresponding row of. At a time, only one of the storage elements in the columns of the encoder array is energized to the distributor so that the distributor can broadcast any row from the input storage array to any row in the output storage array. To accomplish this, the distributor must satisfy three conditions.

1. Center the output beam with respect to its optical axis, then center the decoder row.

2. Disperse the output beam over the entire height of the rows of the decoder array.

3. Control the horizontal dispersion of the output beam and reduce crosstalk between adjacent rows. The first two requirements are satisfied by the Fourier transform performed by distribution. Distributor
The signal light at 103 is simply the light entering the system from the energized optical elements in the decoder array. The only light source in an ideal optical switch is a high-level optical information signal representing, for example, a logical "1" emitted from an energized row of the encoder array. All other optical elements in the encoder array of the ideal system are opaque and do not contribute light to the distributor. As a result, an ideal encoder arrangement requires a device with an infinite contrast ratio. The contrast ratio is the emission of light emitted from the bit of the activated input optical element, the emission of light emitted from the deactivated input optical element bit, and the emission of light from the deactivated input optical element bit. Is the ratio of the emitted light radiation. Unfortunately, commercially available devices and many experimental devices can only achieve contrast ratios of up to 100. Since such contrast ratios are relatively low, a significant amount of unwanted light enters the distributor from the deactivated row of deactivated encoders. This light originates from the background array elements surrounding the energized row and can be referred to as background light. Background light can interfere with the information signal light in the distributor, especially in encoder arrays.
This also results in background noise and noise, as it can cause bit errors in the decoder arrangement, for example, by interfering with a low level optical information signal pair representing a logical "0" from the 103 activated rows. May be called. The optical switch 100 includes
There are two types of background noise. Vertical background noise is background light emitted from optical elements in the same column as the information signal light. Horizontal background noise is background light originating from adjacent columns of optical elements and dispersing across column boundaries. The vertical and horizontal background noise combine to form the overall background of the optical switch.

In order to satisfy the three previously mentioned requirements and to minimize noise problems, the distributor 103 comprises three subsystems: a Fourier transform subsystem 116, a horizontal image enhancement subsystem 117, and a horizontal enhancement system 118.

FIG. 5 shows the subsystems 116, 117 and 118 and the input plane of the distributor system 103.
Encoder array 102 located at 500 and decoder array 104 located at output plane 505 of the distributor system
In addition, a detailed diagram of the distributor 103 is shown. A side view of the distributor 103 and the subsystems of the encoder arrangement 102 and the decoder arrangement 104 of FIG. 5 are shown in FIG. Shown in FIG. 7 is a top view of the apparatus of FIG. Fourier transform subsystem 116
Includes a convergent cylindrical lens 501 labeled L1,
This produces a one-dimensional Fourier transform of the input electric field pattern 550 at its focal plane 502 at the input plane from the optical elements 102 (R1, C8) of the encoder array 102. Since the distance d ₀ from the encoder array 102 to the converging lens 501 is not necessarily equal to the focal length of the lens, the image at the focal plane 502 is actually a Fourier transform of the electric field pattern from the encoder array multiplied by a phase term. Has become. But,
Since the operation of the decoder array is responsive to the magnitude of the light emission, these phase terms can be neglected. Due to the transformation, the spatial frequencies appearing in the image of the encoder array appear as bright areas or spots in this focal plane. Since the input image of the encoder array is a single square opening with the side length given by P,
The electric field distribution pattern in the vertical direction (ie, the y direction) is Given by If If the converging lens 501 has a focal length f _1, the vertical direction of the Fourier transform 55 of the focal plane of the lens
The image of 1 is an electric field approximated by the following equation.

This is the well-known sinc (y) '= sin (πy) /
(Πy) is a vertical function. Since only radiation in this plane is detected, the signal detected is proportional to the square of the Fourier transform (sinc ² ), which is the electric field spectrum of the input image. The power spectrum (and sinc is independent of the vertical position of the square aperture in the encoder array)
The main lobe of pattern ² is always centered on the optical axis of the converging lens 501. Therefore, the first of the three requirements is satisfied. Since most of the energy of the on-state optical element is contained in the main lobe, this is
This is the portion of the output image 552 used to detect the output plane 505 where 104 is located. Unfortunately, the vertical spread of this lobe of the function of sinc ² may not be enough to satisfy the second requirement of the distributor. The first vertical 0 is This is because here It is.

The vertical augmentation subsystem 118 spreads the output beam over the entire height of the columns of the decoder array to satisfy sinc ² to satisfy the requirements of the second distributor.
Expand the vertical spread of the function. Various known optical imaging systems may be used to increase the Fourier distribution in the output plane. Advantageously, a single cylindrical dispersing lens 5 named L3 for the vertical magnification subsystem 118
There is a method in which an image of Fourier transform of the focal plane 502 is formed on an output plane 505 in which a decoder array is arranged. Since the dispersion lens 504 is provided on the left of the Fourier plane,
The Fourier transform image is vertically magnified and appears on the output surface instead of the Fourier surface. The distance to the Fourier transform plane from the dispersion lens expressed in S ₀ (a negative number), Expressed the distance to the image output surface from the dispersion lens S _i, the first of the sinc function, which is enlarged in the output plane at that time The vertical zero point of And this is Will be given by Where F _scaled represents the expanded Fourier transform 522 that occurs on the output surface. Assuming that the axis-parallel approximation is valid, the well-known lens equations and equations describing lens magnification yield two very useful equations for system design.

Here, M _required is the magnification of the system required to spread the main lobe of the sinc function over the entire output plane 505, and f ₂ is the negative focal length for dispersion.

To make the necessary expansion is often the distance S _i increases. This results in horizontal dispersion in the propagation of the beam through the system, and the beam width is approximately , The beam width b may be too large. Where z is the total length of the system. This creates a horizontal noise problem between adjacent columns of the system.

The horizontal imaging subsystem 117 controls the horizontal noise problem by imaging the beam in the horizontal direction as shown in FIG. By refocusing the light in this way, light that may be noisy in the system can be redirected to useful points on the output surface and the level of the output signal can be increased. Horizontal imaging subsystem
117 includes an imaging lens 506, also called L2,
This is a cylindrical convergent lens rotated 90 degrees from the horizontal direction of the other two lenses in FIGS. If the focal length of lens 506 is f ₃ , the distance from the encoder array is given by S ₁ = 2 (f ₃ ), and the distance from the refocus lens to the decoder array is S ₂ = 2 (f ₃ ). Given. This results in the refocusing lens returning all input beams to their initial horizontal width and directing each input beam to a conjugate point in the output plane shown in FIG. Expansion and Contraction of horizontal beams propagating can be carefully controlled by varying the distance S ₁ and S _2. As a result, since the horizontal spread is included, the third condition of the distributor is also satisfied.

In considering the output waveforms resulting from the array of optical elements in the encoder array 102, it is desirable to discuss the output waveforms resulting from a single optical element in the encoder array. The presence of a single optical element in the input plane means that this switch uses an encoder array with an ideal optical device with infinite contrast ratio. Assuming a single element (width = P) is energized in an ideal encoder arrangement, the electric field waveform at the output surface Is approximately given by:

Vertical electric field waveform generated by optical element at input plane Has a rectangular distribution 801 as shown in FIG. By the operation of the Fourier transform lens 501, the electric field waveform pattern in the vertical direction on the output surface is as shown in FIG. Has a sinc distribution 802 described by Also in this case, the image actually detected in the vertical direction of the output has the interval between zeros described by equation (4) and has the form of sinc ² s. It is important to note that the main lobe of the sinc becomes wider as the dimensions of the optical element decrease. It is important to note from equation (8) that the beam given in the horizontal direction of the output plane is correctly approximated by a rectangular distribution of width P.

When all the light elements of the input surface are de-energized, this forms an electric field distribution pattern similar to a square wave when looking down on any row. This rectangular wave distribution can be observed for many encoder arrays because many devices are produced in a rectangular array structure. This rectangular electric field distribution pattern 901 has a noise input pattern as shown in FIG. Corresponding to Therefore, the vertical electric field distribution pattern 902
Is described by the following equation.

Here, P is the size of the optical element, and G is the size of the gap between the elements. Since the encoder actually has a finite spatial extent, this image is actually clipped by the window function w (y). Therefore, the following equation is obtained.

Here, * indicates a convolution operation, and comb (y / (P
+ G)) is a sequence of Dirac delta functions spread over distance P + G. If all rows of the encoder are deactivated using a threshold operation, all elements of the input image will have approximately the same emission. Therefore, the vertical electric field waveform Is generated on the input surface and the vertical electric field waveform pattern is As shown on the output surface. On the output surface, the effect of having multiple light sources emitting on the input surface instead of a single light source element is significant. become that way. If a window of rect (y / w) is applied to the input image and the window length W is much larger than the element size P, the following is obtained.

In the output waveform, the distance from the center to the center is λf ₁ S _i / ((P
+ G) a sequence of sinc pulses of narrow width as S ₀ ), where P is the size of the element and G is the size of the gap between the elements shown in FIG. Each of the sinc pulses has a zero point of λf ₁ s _i /
(Ws ₀ ) interval, and the entire pulse train has a wide sinc
Is Modulated by As the window length W increases (ie, as the size of the encoder array increases), the width of the main lobe of the sinc pulse decreases. Furthermore, as the size of the gap between the elements of the encoder array decreases, the space between the sinc pulse elements on the output surface increases because the rectangular wave image has a higher spatial frequency.

When all elements of the encoder array are deactivated except one, The electric field of the waveform distribution pattern 1001 described in (1) is generated on the input surface as shown in FIG. Therefore, the following equation is established.

Where A is Where k is the contrast ratio of the device in the encoder array. The vertical electric field distribution waveform pattern on the output surface is , And the Fourier transform is a linear operation, so as shown in FIG.

here Is shown by the dashed line 1002, Is shown by the solid line 1003.

FIG. 10 also shows two output regions, where the output signal is sampled by a detector such as an optical device element of a decoder array 104 provided at the output surface. These regions are in the broad sinc main lobe caused by the information signal, but they are limited by the narrow sinc main lobe resulting from the background noise. Thus, the photodetector samples the high emission of the main lobe of the output information signal, while sampling the low emission for the wide lobe of the output noise. When the decoder elements are arranged accordingly, the low emission of the noise signal is negligible compared to the high emission of the information signal.

In order to indicate the characteristics of the operation of the distributor 103, it is effective to define the signal-to-noise ratio (SNR) of the switch. It is clear that the output signal-to-noise ratio changes depending on the sample position on the output surface due to the variation of the electric field distribution waveform pattern on the output surface. As a result, it is of interest to examine the signal-to-noise ratio as a function of the vertical deviation from the optical axis of the distributor. This function is called SNR (y).

Before defining SNR (y), it is necessary to define the case where "noise signal" and "information signal" are applied to the electric field waveform. Background input noise is defined as the electric field waveform present when none of the optical elements of the encoder array 102 are energized. This is shown in FIG. And the resulting output noise corresponds to waveform 901 described by Is represented by the waveform 902 described by As mentioned earlier, this noise is the unwanted light that enters the system through the deactivated encoder element when the bias beam 153 is provided. The Fourier transform nature of the lens is used to force this noise into discrete areas of the output surface.

The input of the system is the "information signal" which is the electric field waveform that exists when a single optical element is turned on in the encoder arrangement and emits an optical signal with a high power level compared to the power level of the noise signal. Is defined as This light is also considered to be part of the input signal because light from the deactivated element is also present when the activated element is known to be on. As a result, as shown in FIG. The waveform 1001 described by represents the input signal, Represents an output signal. This results in:

To enable the use of low contrast ratio devices in the encoder array 102, windowing techniques used in digital signal processing applications can be used. Research on windows and digital signal processing stems from the requirement that all signals must be aborted before applying to digital signal processing techniques. Normal truncation is equivalent to applying a rectangular window to the signal as described above. In the frequency domain, the Fourier transform caused by the truncated signal is the convolution of the Fourier transform of the original uncensored signal and the Fourier transform of the rectangular window. (See Equation 12) Since the Fourier transform of a rectangular window is a sinc function, the sidelobes of the sinc curve tend to disperse the initial signal spectrum distribution. As a result, using a rectangular window tends to increase the bandwidth of the Fourier transform of the original signal. This phenomenon is often referred to as spectral leakage and is observed using any window for truncation, not just rectangular windows. The amount of energy spreading to the side lobes is a function of the window used. The Fourier transform of a rectangular window tends to make the main lobe smaller and the side lobe larger. Sidelobes are small in the Fourier transform of other windows, such as triangular windows, Hamming windows, and glass windows. A window having a low side lobe can be used to improve the signal to noise ratio of the optical switch 100.

To know how the windowing technique is used in optical switch assembly 100, the final height of encoder array 102 relative to the periodic distribution of light emission formed by the elements of the encoder array is described. It is important to know that is effectively effecting the window. A good way to visualize this is to imagine the encoder array 102 extending vertically to infinity. As a result, the switch has an infinite number of rows. In this case, the Fourier transform of an infinite rectangular wave present in the input plane is a sequence of delta functions of the changing light.
Now if a mask is placed before the infinitely long encoder array and blocks all but a few lines,
The input signal is clearly windowed using a rectangular window, and the delta function on the Fourier plane will spread as a result of spectral leakage. If a window other than a rectangular window is used, this spectral leakage can be minimized. The use of Fourier and window techniques in the light domain requires that the system use coherent light from a coherent light source 14. In many lasers, the shape of the emission is Gaussian, so the bias beam 15 is placed on the encoder array formed by the amorphous beamforming system.
Gaussian windows can be easily formed by using a laser with an image as 3.

The effect of applying the Gaussian window to the distributor 103 is important. Applying a Gaussian window provides a dramatic improvement over using a rectangular window in the output region between noise peaks. Such an improvement in signal-to-noise ratio is achieved by a beamforming system 115 that illuminates the encoder array 102.
When illuminating the encoder array with a Gaussian window,
This is a direct result of the significant reduction in side lobes of the noise signal.

Unfortunately, such an improvement in signal-to-noise ratio is not just realized. There is a trade-off that must be considered for this. The first trade-off is that the main lobe of the noise peak in the noise signal becomes wider as the width of the Gaussian beam decreases. As a result, the size of the operating area of the output with a high signal-to-noise ratio tends to decrease, thereby reducing the vertical distance of the output sampling of the decoder array 104.

A second trade-off is that the rows of the encoder array 102 illuminated by the tail of the Gaussian beam do not generate enough light to switch the devices at the output surface. As a result, these input rows cannot be used in the switching operation.
In fact, the center of the Gaussian beam, referred to as the "input operating area", will have an area of available input rows nearby. As the optical element moves from the center of the Gaussian beam toward the tail, the signal to noise ratio decreases. It is important to note that the input operating area is a function of the Gaussian beamwidth in addition to the required signal-to-noise ratio on the output plane.

When all elements of the encoder array are deenergized, some of the light can pass through the encoder array because there is a finite contrast ratio k in a real system.
The encoder arrangement also has a threshold function such that all off elements (both energized and other rows) of the Fourier transform input surface have the same radiation or electric field magnitude. If a linear, non-threshold encoder array is used, problems can occur because both on and off elements can be present in a row of the input storage array that is not energized in the encoder array.
The different emissions from these elements are linear To propagate through a linear encoder array. As a result, there are two different off light emissions for the deactivated information bits. Because of this, the periodicity of the input image is lost and the advantage with respect to the signal-to-noise ratio is lost.

The possibility of such a problem is solved by using a threshold gate in the encoder arrangement. Other solutions are possible by configuring the encoder array with a latching device, such as a three-state symmetric self-electro-optic effect device as described above. Optical switch assembly 10
The zero operation requires that only one row be latched into the encoder array at any given time, as described above.

FIG. 14 shows an input storage array sent from the switch control circuit 110 to the encoder 102 via the bus 172 for selective switching through the distributor 103.
FIG. 7 illustrates three idealized encoder timing control signals 1401-1403 for data words in three different time slots stored in row 102. FIG. The first two lines of information to be switched by the distributor as shown in FIG.
(P-1) is stored in the last row 132RP. This information was previously stored in the input storage array at the end of time frame (F-1). The third row of information to be switched through the distributor is stored in row 132R1 of the first storage array as shown in FIG. This third row of information has been stored during time frame F.
The timing of the distributor system is synchronized by a well-known internal clock of the switch control circuit 110.
13th demand for distributor system timing
The low power level of the optical timing signal 162 as shown between times t6 and t7 in the figure is the optical power level of the encoder control signals 1401-1403 or the decoder control transmitted from the switch control circuit 110 to the decoder array 104 via the bus 173. Signal 140
It does not match the optical power level of 4-1406. Thus, the optical timing control signal must be at a high power level while being switched through the distributor system at each selected row of information in the storage array.

Last input storage array row through distributor
To switch the information in the row before 132R (P-1), the switch control circuit activates the encoder control signal associated with the selected row to electrically activate the devices in that row. At time t1. All other rows of encoder 102 are electrically de-energized because their control signals are maintained at low power levels. While the encoder signal 1401 is at the high power level, the control circuit 110
Operates the coherent light source 114 via the bus 168 so as to generate a low power level during the time t1t2, and a low level pulse is applied to the bias beam 153 incident on all the flip-flop pairs of the encoder 102. Can be
When a low level pulse is applied to the bias beam 153, the timing control signal 162 incident on all the flip-flop pairs of the input storage array 132 goes to a high power level, so that all data stored in the input storage array Is
From there, the encoder array through the information signal pair 192 (P, M)
Propagates to 102, but the information is
Go only to (P-1). Since data is accumulated in the energized encoder row, the bias beam 153 is set at the time t.
Return to high power level at 2 and activate encoder array 102
The information stored in the row of the distributor system 10
Through 3, go to the decoder array 104. Also, at time t2, the decoder control signal 1404 assumes a high logic level (1) and electrically energizes the elements of one or more rows of the decoder to receive a pair of optical information signals. During the time interval t2-t3, the clocked light source is controlled by the switch control circuit 110 through the bus 169 to generate the bias beam 1
Pulse 54 to a low power level and optically energize the elements of the decoder array to receive the pair of optical information signals through the distributor. As a result, the information represented by the information signal pair passing through distributor 103 is stored in the flip-flop pairs of the electrically and optically activated rows of the decoder array.

After the data has been stored in the energized 104 decoder array rows, at time t3, the light source 145 pulses the beam 154 to a high power level and the row of information now stored in the decoder array 104. Between the time t3t4 and the output system
Propagation is made to the output storage array 133 of 105. Bus176
The light source 142, which is controlled by the switch control circuit 110 via the
As it remains incident on 133, the information of the activated decoder row is stored in the corresponding row of the output storage array. An information signal is stored in the output storage array between times t3 and t4 when control timing signal 154 becomes a high power level pulse on decoder array 104. All other rows of the decoder array are disabled and emit a pair of optical signals that are both low. As a result, these low power level optical signal pairs will not affect the rows of the storage array of the output at which they are incident. At time t4, the electrical decoder control signal returns to a low logic level (0) and the light control signal 154 also returns to a low power level, thereby causing the decoder array row 104 (P-1).
Is electrically and optically deactivated.

At time t3, the encoder bias beam 153 becomes a low power level pulse and the encoder control signal 1402
Goes high and the last encoder row 102RP is electrically energized, the sequence described above is repeated. As a result, the corresponding input storage array row 132RP
Are transferred in parallel to start the next cycle of the row switching process. At time t4, when the bias beam 153 of the encoder goes to a high power level,
The new row of information represented by the information signal pair propagates through the distributor to the decoder array. At the same time, the decoder control signal 1405 changes to a higher power level to activate a new decoder row, and the decoder bias beam 154 changes to a lower power level to decode information from the activated encoder array row. Energize to memorize. At time t5, encoder signal 1402 changes to a low logic level, electrically deenergizing encoder array 102RP, and encoder beam 154 returns to a high power level to pass the last row of data from the decoder to the output storage array. Let it. At time t6, the decoder signal 1405 and the bias beam
154 goes low, causing the next row of data from the distributor to be stored in the decoder.

After all rows of information of the time frame (F-1) of the input storage array have been transferred to the output storage array 133, the above-described process is performed for the first row of information of the time frame F in the first storage array. Is repeated. This involves an encoder control signal 1403 and a decoder control signal 1406, as shown in FIG.

Like the encoder array 102, the decoder array 104
As shown, it includes a plurality of optical storage elements, such as three-state symmetric SEEDs arranged in rows 104R1-104RP and columns 104C10-104CM. The switch control circuit 110 via the bus 173
One or more rows of the decoder array are activated to store the information represented by the information signal pairs distributed throughout the corresponding column pairs. Decoder array 104
After the information is stored in the activated row of the well-known polarization beam splitter 119 inserted between the distributor system 103 and the decoder array 104, a clock such as a bias beam 154 from a clocked light source 145 is applied. Is applied to each of the devices of the previously described decoder arrangement. The size of the clocked bias beam is uniform. The bias beam is provided from a coherent or non-coherent light source to the entire decoder array as previously described in FIG. Only the activated rows of the decoder array output the information contained in the output system 105.
To generate a complementary pair of optical signals having a high power level and a low power level for storage. Rows of deactivated devices as shown in Table A emit only low power level pairs. As noted in previously cited U.S. Pat.No. 4,754,132,
In order to change the state of the symmetric SEED, it is necessary that a ratio between the powers of the two optical control signal beams is higher than a predetermined value. Since the power level of the light output signal pair from each deactivated element is the same, the light output signal pair is the same and the light output signal pair changes the state of the symmetric SEED storage element of output system 105. Never. However, the ratio of the high power level and low power level output optical signals representing the information of each of the activated decoder elements is sufficiently large to change the state of the storage element in the output system. Thus, only information from the activated decoder row will be stored in the output system.

FIG. 11 shows an output storage array 133 for storing the rows of information selected by the decoder array 104 and an associated output pair 160 from the spatially separated form of the stored information.
(1) A detailed view of an optical output system including an output system register unit 129 for converting to a time separated format for serial transmission to -160 (N). An output system interface unit 127, 128 including an output storage array 133 and an output shift register unit 129 is sandwiched between optically transparent material 126.

Output storage interface unit 127 is decoder array 1
It includes a well-known imaging system 120 for forming an image of the optical signal pair 194 (P, n) from the elements of 04 through the beam splitter 121 onto a receiving surface having a corresponding position in the output storage array 133. Well-known polarizing beam splitter 1
21 is an optical bias beam 156 with a clock from the light source 142
To each of the optical storage elements of the output storage array 133.

As with the input storage array 101, the output storage array 133 has a row 133R
1-133RP and includes a plurality of optical storage elements such as symmetric SEEDs arranged periodically as columns 133C1-133C8.
In response to the bias beam 156, the output storage array 133 emits a pair of complementary optical information signals from each of its storage elements representing the information contained therein for storage in the output shift register unit 129. After information from a certain time frame period is stored in the output shift register unit,
The output storage array stores the information stored in the shift register unit in series with the output fiber pair 16 while storing the next frame of switched information from the decoder array 104.
It is ready to shift out to 0 (1) -161 (N) in series.

The optical shift register interface unit 128 is inserted between the output storage array 133 and the shift register 129 to form an image of the optical signal from the storage array on the corresponding master storage element of the output shift register unit, and the optical timing control signal 157-1. 159 is provided to the elements of the shift register unit. The switch control unit 110 controls the spatial light modulator 141 via the bus 174, and emits an optical timing control signal 157-159 in response to an optical power signal 155 from a light source (not shown). The shift register interface unit includes an imaging system 122 that forms an image of the optical signal from each storage element into a master SEED having a corresponding location in each of the shift register units. A well-known polarizing beam splitter 133 directs a timing signal to each of the storage elements of the shift register unit 129.

Similarly to the input shift register unit 131, the output shift register unit 129 includes a plurality of optical shift register rows 129.
R1-129RP and serially output information stored in a spatially separated form to the pair of optical fibers 160 (1) -160.
(N) is shifted serially to the relevant pair and transmitted in a temporally separated form. Output shift register unit 12
9 is an optical storage element array 134 arranged in rows 134R1-134RP and columns 134C1-134C8, rows 125R1J-125RP and columns 125C1-125.
Includes a similar array of holograms 125 arranged in C8. An optically transparent spacer material 124 is interposed between the storage element and the hologram array. Each column includes a pair of storage elements designated as master and slave for storing and shifting bits of information. Arrangement 12 of corresponding pairs of reflection holograms placed facing storage element pairs
The pairs are interconnected such that 5 is a well-known master-slave flip-flop configuration. The pairs of storage elements in each row are also optically interconnected by facing holograms to shift complementary optical information signals from one storage pair to another. For example, the first master-slave pair of storage elements in row 134R1 is 134 (R1, C8) M, 134 (R1,
C8) Designated as S. Opposing holograms are similarly named as shown.

FIG. 15 shows an idealized optical timing control signal 157-159, a clocked optical bias beam.
FIG. 5 is a timing diagram for the time t of 156 and the complementary optical information signal pair on output fiber pair 160 (1). As shown, complementary optical information signal pair 1502 represents the data word of the last time slot, which is the time frame (F-2).
8 bits of information B1-B8. Similar pair 1503 represents the first few bits of the first time slot of time frame (F-1). The information from the time frame (F-2) is the time frame (F-3) between the times t3 and t4.
Bias beam 1 at the end
Output storage array 13 when 56 becomes high power level pulse
3 to the output shift register 129.
During this time, the electronically controlled spatial light modulator 141 keeps the timing control signals 157, 158 and 159 at a low power level, so that the data represented by the information signal pair is transferred from the output storage array 133 to the output shift register unit 129. Propagate to As a result, the data is stored in the flip-flop pair of the output shift register unit 129. For example, when the data in rows 129R1-129R3 is stored in the output shift register unit, it is shifted out one row at a time to a single output fiber pair 160 (1). Bias beam 159 is maintained at a low power level for all output shift register rows that are not transferring information out the output fiber pair.

Output shift register unit 129 includes holograms arranged in rows 125R1-125RP and columns 125C1-125C8, as well as array 134 of optical storage elements arranged in rows 134R1-134RP and columns 134C1-134C8. An optically transparent material 124 is interposed between the storage element and the hologram array. Each column of the array 134 includes master and slave optical storage elements that store and shift bits of information. A corresponding pair of holograms in array 125 facing the pair of storage elements optically interconnects the pair of storage elements, creating a well-known master-slave flip-flop configuration. The pairs of storage elements in each row are also optically interconnected by hologram pairs to form rows of optical shift registers that shift pairs of complementary optical signals from one pair of storage elements. For example, the first master-slave pair in row 134R1 of storage elements may have 134 (R1, C8) M and 134 (R1,
C1) S, the last pair in the row is 134 (R1, C8) M and 13
4 (R1, C8) S. The holograms facing this have the same numbers as shown.

The shift of the 8-bit information B1-B8 stored in the output shift register array 129 is performed by the input shift register unit 131.
Is performed in the same manner as described above. Bit B1
Is stored in the master flip-flop storage element 134 (R3, C8) M, and the other bits B2-B8 are stored in the other master flip-flop storage elements of the row 134R3. Provides the timing G signal 58 to all master flip-flop storage elements in that row,
Timing signal 157 must be provided to all slave flip-flop storage elements in the same row. As shown in FIG. 15, at time t0, the bias beam becomes a high-level pulse only in the row where the complementary information signal is shifted out to the output fiber pair. High power level bias beam 158 for master flip-flop storage element
, The information signal pair is emitted from the master flip-flop storage element and directed to the opposing hologram of array 125, where the signal is again provided to the paired adjacent flip-flop storage element. Since the bias beam 157 is at a low level between times t0 and t1, the information on the master flip-flop storage element is transferred to the flave flip-flop storage element. Further, the information signal from master flip-flop storage element 134 (R3, C8) M goes to hologram 125 (R3, C8) M, which directs this signal to output fiber pair 160 (1).

After the transfer of the information ends at time t1, the timing control signal 158 changes to a low power level. As a result of this,
A complementary information signal pair representing the information stored in the slave flip-flop storage element is provided to the facing hologram, which returns the signal to the adjacent master flip-flop storage element. This means that data has been transferred from the slave flip-flop storage element to the master flip-flop storage element. At time t2, the timing signal
157 returns to a low power level and timing signal 158 returns to a high power level. The above sequence is repeated one after the other to shift each bit of hand information through the rows of the shift register, resulting in an output fiber power of 160 (1). Thus, the desired parallel-to-serial conversion is performed to convert the spatially separated information into a temporally separated format. As shown in FIG. 13, a time frame boundary exists at time t8. FIG. 14 shows that the boundary of the time frame occurs at time t7,
In FIG. 15, the time boundary occurs at time t4. Since the data is switched through the distributor while the time frame F is in the input shift register unit 121 as shown in FIG. 13, the data shifted out after the frame boundary in FIG. Corresponds to the time frame (F-1).

A group of rows of output shift registers, such as rows 129R1-129R3, each containing an information time slot, is 160 (1).
Associated with each output fiber pair. This group contains the time frame information for serially transmitting the time separated information to the associated output fiber pair. The control timing signals 156-159 are clocked at the data rate of the output fiber transmission system and release information signals from the last device to each of the activated shift register rows. Each group of rows is automatically activated in a known manner, or alternatively is controlled by output fiber system 190 through control bus 174.
In order to synchronize the optical switch with the output optical carrier 190, a well-known timing control signal is output from the output control circuit 110 to the bus 167.
Sent to

Output shift register unit 129 and output fiber pair 160
Between (1) -160 (N) is a layer of optically transparent spacer material 126. Spacer material 126 provides space for each hologram pair to direct an optical signal pair to the associated output fiber pair with minimal transmission loss.

In summary, the optical switch assembly 100 performs the operation of the space division optical switch assembly and the time division optical switch assembly in the optical domain completely within the stage of the switch. This is performed by a distributor system that spatially distributes the optical information signal pairs from encoder array 102 to all rows of decoder array 124. The optical information received in series is stored in a row of the input shift register unit 131. In order to take advantage of the parallel operation of this system, the data to be exchanged must be provided to the system in a parallel format. This is performed by the input shift register unit which shifts the received data serially for the row associated with the particular data time slot. Similarly, the opposite operation is performed by the output shift register unit.
129, in which output information exchanged through the distributor 103 is read out serially from a row of output shift registers and output fiber pairs 160 (1) -160.
(N). The shift operation of the optical information, which becomes a shift register unit, allows the distributor to operate at a much lower speed than the incoming or outgoing data. Only the input and output storage units are required to operate at the rate of the input and output serial bits. The input system 101 converts the bit stream of the temporally separated optical signal into a spatially separated bit pattern. The information is then output through a distributor in parallel
Switched to 105 and switch control circuit 1
An encoder arrangement 102 and a decoder arrangement 104 are used under the control of 10. Thus, data in any time slot on the optical fiber can be switched to any other time slot in the output fiber, thereby performing both a time switch and a space switch.

As described above, the optical switch assembly of the present invention has been described in detail. I want to summarize.

Principle of Operation The optical switch assembly of the present invention includes an optical distributor (103) that directs light from the array of optical transmitter elements (102) to the array of optical receiver elements (104).
These elements are arranged in rows extending in the x direction and columns extending in the y direction.
They are arranged in a dimension. Importantly, these elements can be selectively activated (energized) row by row. The distributor is made with focusing means (501, 504, 506) having different focusing characteristics in the x and y directions.
Due to the different focusing characteristics, light from any element in any column of the array of transmitter elements is distributed across all elements in the corresponding row of the array of receiver elements. as a result,
Light from any row of transmitter elements is selectively transferred to any one or more of the rows of receiver elements.
The optical switch assembly also includes an optical input system (101) comprising an optical shift register unit (131) configured to receive optical signals from the plurality of serial optical signal input channels (150) and an input storage element array. In. The optical input system spatially separates the serial optical signals. An input storage element array (132) receives and stores the spatially separated signals and applies them to an array of transmitter elements. Each of the input channels is uniquely assigned to one or more rows of the input storage element array. One or more input channels are assigned to a plurality of rows of the input storage array corresponding to different time slots.

Further, the optical switch of the present invention further comprises an output storage element array for storing signals from the array of receiver elements.
And an optical output system (105) comprising an output shift register unit (129) configured to serially shift a signal stored in the output storage element array to a plurality of serial optical signal output channels (160). I have. Each of the output channels is uniquely assigned to one or more rows of the output storage element array.

The optical switch operates to perform both space division switching and time division switching of optical signals within a single switching stage. An input fiber may have several time slots. Each time slot of each input fiber is assigned to a different row in the array of transmitter elements. The information in one transmitter row is transferred to one or more optional receiver rows for distributor operation. Therefore, it is possible to exchange and switch the time slot of one input fiber to any time slot on any output fiber.

The distributor advantageously distributes the optical signals of the individual columns from the selected row of the array of input locations centered with respect to its optical axis and the center of the corresponding columns of the output array, and the individual corresponding output Any position in the row can be illuminated. The same result is obtained regardless of the vertical position of the optical signal in the input train. Similar results are obtained for each signal emitted from the other columns of the row of the selected input array.

The optical switch assembly of the present invention comprises: an array of optical transmitter elements (102); an array of optical receiver elements (104); An optical distributor (103) and an optical source (150 (1), 151) for directing light through the array of transmitter elements to the array of receiver elements.
(1)) is a main component.

The array of optical transmitter elements 102 comprises an encoding array of individual light beams, ie, an input array. The array 104 of optical receiver elements comprises a decoder array having elements arranged to correspond to the elements in the array 102 (see also FIG. 1). Light source 150 (1) to 150 (N)
Provides light that is directed through the array of transmitter elements to the array of receiver elements. The elements of each array can be selectively activated, or activated, row by row. As can be seen with reference to FIG. 5, the distributor is provided by focusing means 501, 506, 504 having different focusing characteristics in the x and y directions (see the x, y, z coordinate system at the upper left end of FIG. 5). Become. An array of transmitter and receiver elements is located in each conjugate plane of the distributor with respect to focusing in the x-direction. That is,
Referring to FIG. 7, the light input on the leftmost element in the array of transmitter elements 102 is directed, ie, focused, by a lens 506 so that the corresponding optical element is in the rightmost column of the array of receiver elements 104. Will appear. This left-to-right and right-to-left replacement
It is defined as “having a conjugate plane relationship with respect to focusing in the x direction”. FIG. 6 illustrates light transfer in the y-direction (vertical direction). An array of receiver elements is located at the focal plane of the distributor with respect to y-direction focusing. This allows light to be distributed from any of the elements in any row of one column in the array of transmitter elements to all elements in the array of elements in the receiver array. Referring to FIG. 5, the optical element 5 in the top row and the leftmost column of the transmitter array 102 is shown.
50 is mapped in both the x and y directions of the distributor to result in a corresponding light distribution as shown on the receiver array 104 as shown in FIG. That is,
In the x-direction, a conjugate or left-to-right mapping results in a light beam transfer from the leftmost column to the rightmost column. And in the y direction, the input beam is spread to cover all rows in the column as determined by the x direction mapping. In this manner, light from any row of transmitter elements will be selectively transferred to any one or more of the rows of receiver elements. The optical switch assembly of the present invention has the effect of allowing spatial orientation of the information carried by the light beam. In the example we invented earlier,
Referring to FIG. 5, the light beam 550 can be received by any of the corresponding receiving elements arranged in a row (in which the light form 552 is distributed). For a given time slot, any receiver element in the corresponding column can receive information 550. And typically, since only one receiver element in one row of this column is activated (energized), the top row in the transmitter element array 102, from the left column to the rightmost column in the receiver array 104 1
Spatial routing to one row will be done. The optical distributor included in the optical switch assembly of the present invention is an optical device that captures output bits from any row of the input device array and directs a beam of light by the optical distributor. However, since the light beam from each of the input bits must not interfere with the others (Clochtalk must be minimized), each of the bits that are spatially separated across rows must be stored by optics. No. Thus, the distributor functions to spread the light beam in the vertical y-direction and control the divergence of the light beam in the horizontal x-direction. By placing the array in the conjugate plane of the distributor with respect to the x-direction, imaging is performed to control the horizontal dispersion of the beam. On the other hand, the array is positioned at the focal plane of the distributor with respect to the y-direction, causing a Fourier transform and obtaining a vertical spread of the beam.

The present invention overcomes the problems in the prior art by using a single stage light distributor to provide both time and space switching.
This distributor is less complex than the hardware requirements for free space crossbar switches and lithium niobium multistage switches. In the present invention, the same vertical beam spreading effect is achieved regardless of the vertical position of the optical input signal due to the operation of the Fourier transform. This effect can be easily understood by considering that all Fourier transforms of the time function representing signals in the time domain are symmetric with respect to the W = 0 point in the frequency domain. Similarly, the Fourier transform of the input spatial function in the present invention is also symmetric with respect to the optical axis. The time shift of the time signal with respect to the time domain function only results in a phase shift of the Fourier transform. Therefore, the amplitude function of the Fourier transform is not affected at all by the time shift of the time signal. Similarly, the lens in the embodiment of the present invention is
The vertical spatial shift of the input spot does not change the amplitude function of the Fourier transform produced by the output device array, as it causes a Fourier transform of the signal's electric field distribution along the vertical y-axis of the input device array. That is, only the phase of the optical signal reaching the output device array changes. Because the photodetectors of the output device array are not sensitive to the phase of the light, the output detector will receive the same signal regardless of where the signal was activated in the input device array. Activation of the transmitter in the input device array directly reflects the input electric field pattern. Since the transmitter is modulated based on the state of the bits in the word being activated on the bus, this electric field pattern will effectively be formed by the state of the bits in the word.

Another aspect of the invention is the use of a shift register to convert the serial data stream at the input to W bits in a parallel configuration (which can be injected into the distributor). Any shift register has two basic elements: a flip-flop storage element, and means for connecting the output of flip-flop P to the input of flip-flop (P + 1). In the illustrated embodiment, the flip-flop storage element is provided by a SEED device. Also,
The means for connecting the output of one flip-flop to the input of another flip-flop is provided by a hologram that routes the light beam emanating from one SEED device to the input of another SEED device. The hologram is a reflection hologram that reflects a beam in the i-th row of the SEED device and the P-stage to the (P + 1) -th row in the SEED device.

A beam-splitter (beam splitter) 109 is used to guide the optical input signal from the fiber as well as the optical control (ie, clock) signal from the Semtex spatial light modulator onto the input register unit 131. This is because the light from the fiber is P-polarized so that it passes through the beam splitter and travels to the device array, and the light is S-polarized so that the light from the semtex is reflected off the device array. It is accomplished by having The device array to which these signals are directed is a SEED device that operates as a flip-flop in the present invention. The arrangement provides the many flip-flops necessary to implement a shift register as described above.

As described above, by applying Fourier optics, the light beam can be expanded in the vertical direction, so that the intensity distribution generated on the output surface can be concentrated on the optical axis regardless of the vertical position of the input beam.

[Brief description of the drawings]

1 is a diagram of the optical switch assembly of the present invention; FIG. 2 is a detailed illustration of the input system of the optical switch assembly of FIG. 1; FIG. 3 is beam forming of the optical switch assembly of FIG. FIG. 4 is a detailed diagram showing the magnitude of an optical bias beam having a Gaussian distribution formed by the system; FIG. 4 is a circuit diagram of a self-electro-optical effect device used in the encoder array of the optical switch assembly of FIG. 1; 5 is a detailed view of the distributor of the optical switch assembly of FIG. 1; FIG. 6 is a side view of the distributor of FIG. 5; FIG. 7 is a top view of the distributor of FIG. FIG. 9 shows an example of the waveform of the input electric field of a single encoder element of the optical switch assembly and the resulting Fourier transform present in the decoder arrangement; FIG. 10 illustrates exemplary waveforms from the optical elements of a row of the decoder array of the optical switch and the resulting Fourier transform present in the decoder array; FIG. 10 shows noise from the row of the encoder array of the optical switch assembly of FIG. 1; FIG. 11 shows an example waveform and a single optical information signal and the resulting Fourier transform on the decoder array; FIG. 11 is a detailed view of the output system of the optical switch assembly of FIG. 1; FIG. 12 is FIG. 13 to 15 show various optical information signals and optical and electrical timings provided to the input system 101, distributor 103 and output system 105 of the optical switch assembly of FIG. FIG. 4 is a timing chart of a control signal. [Description of Signs of Main Parts] Distributor Means 103 Decoder Means 104 Control Means 110 Encoder Means 102

 ──────────────────────────────────────────────────続 き Continued on front page (72) Inventor Harvard S. Hinton United States 60565 Illinois, Naperville, Calico Avenue 1229 (72) Inventor Frederick Bee. McCormick, Jr. United States 60532 Illinois, Wristle, Apartment 4M, A Bay Drive 5650

Claims (7)

(57) [Claims] An array of optical transmitter elements (102), an array of optical receiver elements (104), and an optical distributor (103) for directing light from the array of transmitter elements to the array of receiver elements. ) Wherein the array is a two-dimensional array, wherein the elements have rows extending in the x-direction and y
Light sources (150 (1) through 150 (1)) for directing light through the array of transmitter elements to the array of receiver elements.
50 (N)), wherein the elements of the array are selectively activatable row by row, wherein the array is positioned at a respective conjugate plane of the distributor with respect to focusing in the x-direction and receives. Focusing means (50) having different focusing characteristics in the x and y directions such that the arrangement of the detector elements is located at the focal plane of the distributor with respect to the focusing in the y direction.
1,504,506) and light from any element in any column of the array of transmitter elements is distributed over all corresponding elements of the array of receiver elements, thereby resulting in any row of transmitter elements. An optical switch assembly characterized in that light from the optical switch assembly is selectively transferred to any one or more of the rows of receiver elements. 2. The optical switch assembly of claim 1, wherein the optical switch assembly is configured to receive optical signals from a plurality of serial optical signal input channels (150) and spatially separate the serial optical signals. An optical shift register unit (131) for receiving and storing the spatially separated signals and corresponding to the rows and columns of the array of transmitter elements to apply them to the array of transmitter elements An optical switch assembly comprising an optical input system (101) comprising an array of input storage elements (132) arranged in rows and columns. 3. The optical switch assembly according to claim 2, wherein each of said input channels is uniquely assigned to one or more rows of an input storage element array. An optical switch assembly, comprising: 4. The optical switch assembly according to claim 3, wherein one or more of said input channels correspond to a plurality of rows of said input storage array and different time slots. An optical switch assembly, wherein the optical switch assembly is assigned to a plurality of rows. 5. The optical switch assembly according to claim 1, wherein the signal from the array of receiver elements is stored for storing signals from the array of receiver elements. An output storage element array (133) configured in rows and columns corresponding to the rows and columns of the array of receiver elements, and a signal stored in the output storage element array to a plurality of serial optical signal output channels (160) An optical switch assembly comprising an optical output system (105) consisting of an output shift register unit (129) configured to shift in series with the optical switch assembly (105). 6. The optical switch assembly according to claim 5, wherein each of said output channels is uniquely assigned to one or more rows of an output storage element array. An optical switch assembly, comprising: 7. The optical switch assembly according to claim 6, wherein one or more of said output channels are a plurality of rows of an output storage element array and correspond to different time slots. An optical switch assembly, wherein the optical switch assembly is assigned to rows.